Price forecasting for U.S. cattle feeders: which technique to apply?

Abstract

The price uncertainty associated with cattle feeding is the catalyst of a volatile economic environment that has exposed U.S. cattle feeders to increased risk levels, thus increasing a need for more accurate price forecasting techniques. This price uncertainty faced by cattle feeders is concentrated in the two main inputs, corn and feeder cattle, and one output, fed cattle. The interrelationship of these components forces cattle feeders to constantly adhere to the basic cattle feeding formula: minimizing both feeder cattle costs and corn costs, and maximizing fed cattle prices. This research strives to evaluate the accuracy of six distinct price forecasting techniques over an eleven year period. The forecast techniques selected for this analysisare the following:1. FAPRI3. AO S5. Univariate Time Series7. Composite 2. WASDE4. Futures Market6. Multivariate Time Series The characteristics of each of the aforementioned forecast techniques are explained within the appropriate chapter. Furthermore, it should be noted that all forecasts from FAPRI, WASDE, and AOS were obtained through their respective published material. The author constructed the futures market forecast by the simple average of certain futures contract average settlement prices. In addition, the author constructed the univariate time series model which provided the univariate forecasts. The multivariate forecasts are obtained from a cointegrated, error-correction multivariate time series model. The composite forecast is the simple average of forecasts for the respective period. The results of twelve forecast comparisons provide an accuracy measurement by the mean squared error (MSE) and the mean absolute percentage error (MAPE). Results from this research conclude that the univariate time series forecast is the most accurate technique overall among the twelve forecast competitions.